214 
TEE SCIENCE OF FRONTIER DELIMITATION. 
carefully, and doing the work which must be done all over again when 
the boundary is delimited. If a reconnaissance is impossible, it is very 
useful to send on a skilled man to fix flags on suitable positions. By 
doing this, all risk of error in recognising points observed is removed. 
But even this precaution is very often difficult to carry out. In my 
experience it was seldom successful. The rate of progress in a densely 
covered country is so slow, the difficulties of cutting a way so great, 
and the risk of going astray so serious that we rarely found that the 
man sent on reached the spot he was sent to, and fixed his flag in time 
to be of service. 
The first step in the work of triangulating is to measure a base as 
far as possible at right angles to the general direction of the frontier. 
This should be carefully done by men trained to the work, as it will 
rarely be found that there is open, level ground in a position to be 
available. But, even if the measurement of the base is not correctly 
done,, the value of the work is not destroyed, provided that the trian¬ 
gulation is consistently carried out, and that the opening and closing 
points are fixed, for the error will be constant, and all that is necessary 
is to fit the triangulation on to the fixed points and to add or substract 
a percentage to or from each computed side and the base. 
After the base has been measured, an observation should be taken 
for finding its true bearing, and at the same time the magnetic 
deviation of the compasses used by the commission should be found. 
The latitude and longitude of one end of the base line should be 
fixed. If the position of the base is such that it is not easily possible 
to do this., it will be sufficient to connect the end with the point which 
has been fixed in latitude and longitude by making it a trigonometrical 
point. 
The most important and at the same time the most difficult part of 
triangulation in rough country, is to recognise the points which have 
been observed. Failure to do this will lead to the survey breaking 
down, or to the introduction of serious errors. The observing party 
must in the first instance, in entering the round of angles in the field- 
book, describe both in writing and by a rough sketch each new point 
when first observed. In addition to this it is useful to set off the 
angles as observed on a plane table, and to make a rough sketch of 
each point at the end of the ray directed on it. But the best pre¬ 
caution, according to my experience, is to take a careful magnetic 
bearing of the R.O, at each trigonometrical station. From the observed 
angles the bearing of each point observed is easily ascertained, and on 
advancing to these stations, the back bearing to the station from which 
they have been observed can be found. For example, you observe a 
point, and enter in your book (with sketch, “ withered single tree just 
below crest of conical hill”). You proceed to this point. If you 
could keep your eyes on it the whole time you were moving, it would 
be easy to recognize it; but what actually happens is this : you drop 
into a densely covered valley and lose sight of your point, and you 
move along this valley for some distance, or perhaps cross several 
ridges without being able to get a view of it. At length you arrive in 
